Claude lifts a Riemann-zero bound to 67.2%

An unreleased research version of Claude improved the known lower bound for the share of Riemann-zeta zeros on the critical line from 41.6% to 67.2%, according to Anthropic’s account of the result. The model spent 31 million output tokens across two Claude Code sessions, and Anthropic researchers validated the paper while a Lean formalization passes the project’s stated checks; the work does not prove the Riemann hypothesis itself. Scientific-AI teams following GenScript’s four-day wet-lab loop should file the pattern: model-generated conjectures become useful research only when domain experts and machine-checkable artifacts can interrogate them.